Modeling 2D vs. 3D Flow in a Circular Pipe

[Mad_Jack]

Hi everyone!

I have the following problem. I have a two phase flow in a circular pipe and I want to model it. I need to decide and justify whether the model should be 2D or 3D. For the moment, I can say only one thing: considering z-axis in the length direction (that it is the dominant one) I have to neglect x or y. If I do this, I suppose I would loose the circularity of the pipe because I consider x or y as the elongated direction, obtaining something more similar to the flow between two flat plates. Then, probably, I need a 3D approach.

What do you think about it? Do you have any suggestions? What other things could I loose if I choose a 2D model?

Thank you very much!

 

[anorlunda]

Is there any chance of even partial separation of the phases because of gravity?

 

[boneh3ad]

Are you familiar with the concept of axisymmetry?

 

[Mad_Jack]

Sorry, I forgot to mention that the problem is in absence of gravity and the two phases won't be separated because the diameter is beneath the critical one.

Yes, I'm quite familiar with axisimmetry.

 

[boneh3ad]

Then I suppose the next question is, if you are familiar with axisymmetry, why are you trying to work this in Cartesian coordinates? Your question shouldn't be whether or not you can neglect $x$ or $y$. Whether the flow is 2D or not you'll lose information doing that. The real question you should ask is whether this is axisymmetric.

 

[Mad_Jack]

I'm not trying to work it in cartesian coordinates, it was just a way to explain my thoughts on what I was supposed to solve using a 2D model as I told. But then, if I consider the channel as planar, cartesian coordinates should work well.

In any case the problem is not axysimmetric.

 

[boneh3ad]

You can't consider the channel as planar, though. You said in the original post that you are dealing with a circular pipe, so treating the problem in cylindrical coordinates is appropriate. I brought up axisymmetry because the operative question regarding 2D vs. 3D flow here is therefore whether or not the $\theta$ direction can be neglected, and that depends on the situation and the level of fidelity you require.

 

[Mad_Jack]

Thank you very much Bone! You helped me a lot to gain a deeper insight in my problem. In the next few days I'll see if it works! ;-)